Assessment of the Dependence of Renewable Penetration Level in a Weakly Connected System on Available Control and Defence Systems

Emanuele Ciapessoni et al. Int. Journal of Engineering Research and Application www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 10, ( Part -1) October 2016, pp.11-19
www.ijera.com 11 | P a g e
Assessment of the Dependence of Renewable Penetration Level in
a Weakly Connected System on Available Control and Defence
Systems
Emanuele Ciapessoni*, Diego Cirio*, Antonio Gatti*, Andrea Pitto*
*(Ricerca sul Sistema Energetico - RSE S.p.A., Milan, Italy)
ABSTRACT
The integration of non-dispatchable renewable energy sources (RES) poses serious challenges to power system
planning and operation. At system level, a major issue regards the short and long-term frequency stability due to
the reduction of regulating reserve margins, in turn brought about by the displacement of conventional
generation and lack of contribution by the RES. This problem is especially present in isolated or weakly
connected systems, and may become the limiting factor for renewable penetration.
The paper proposes a fast algorithm to quantify the benefits of advanced control, protection and defense systems
in increasing RES penetration in weakly connected power systems. In fact, advanced solutions relying on both
system design and operating measures may be of paramount importance to allow the secure integration of more
RES generation. The algorithm is tested on a 2020 scenario of the Sardinian power system and it is cross-
validated against time domain simulation.
Keywords: frequency, stability, power systems, control, defense, protection, wind power, photovoltaic
generation
I. INTRODUCTION
The penetration of growing amounts of
non-dispatchable Renewable Energy Sources (RES),
mainly wind and solar (photovoltaic, PV), allows to
reduce the consumption share covered by fossil-
fuelled power plants [1]-[3], but introduces a number
of criticalities in power system planning and
operation [4]. In fact, RES are often located far from
the bulk transmission system, therefore congestion
and stability problems may arise. The network must
be developed to connect renewable energy
production and to access flexibility resources such as
reserve capacity and remote storages. Very large
excursions between minimum and maximum values
of RES production call for the availability of
conventional generation just to compensate for
periods of low renewable production [5], [6].
Actually, conventional generation is more and more
called to provide ancillary services rather than
energy services, i.e. to compensate for RES and load
variations rather than cover base load.
Significant issues are associated with
frequency regulation. Renewable plants are usually
operated with no upward margin in order not to
“waste” the primary resource, thus they do not
contribute to under frequency regulation. Moreover,
the downward regulation required by grid codes is
often less demanding than for conventional units.
Because RES have priority of dispatch (e.g. in
Europe), the “displacement” of conventional
generation leads to a decrease in online reserve,
unless more stringent constraints are imposed by the
grid codes [7]-[12]. At the same time, forecast
uncertainty of RES generation increases the reserve
requirements from conventional units for balancing
purposes.
RES and more in general Distributed
Generation (DG) connected to the distribution level
may cause local problems, such as power flow
inversion possibly implying protection misoperation,
unintended island, voltage deviations from nominal
values, congestions. However, high penetration of
DG may impact security at system level too. As far
as system defense is concerned, the presence of DG
may substantially affect the effectiveness of the load
shedding schemes, which were designed when DG
was negligible. Restoration process may be
complicated by DG and in general by uncontrollable
generation. In addition, the settings of under/over
frequency relays of DG are, or have been until
recently, very strict in many countries (e.g. 300
mHz in Italy, +200 mHz in Germany): this may
cause the tripping of distributed generation in case of
system disturbances, especially in countries with
large penetration of photovoltaic units in distribution
networks, like Germany and Italy which have solved
this issue by changing the connection requirements
and carrying out retrofitting campaigns to modify
the protection settings [13]. Operation issues related
with PV systems have also been addressed by the
European TSOs [14]-[17]. Frequency problems are
of paramount importance for islanded or weakly
interconnected power systems where larger
frequency deviations are more likely to occur with
respect to continent-scale power systems.
RESEARCH ARTICLE OPEN ACCESS

Different studies in literature tackle aspects
of frequency stability assessment and control in
weakly connected or isolated power systems: in [18]
some issues related to frequency regulation and
primary frequency control models in the Irish system
are addressed. In [19] the -d/dt plane is exploited
to explore new under frequency load-shedding
schemes (UFLS) for isolated power systems by
identifying the frequency stability boundaries. In
[20], [21] dynamic simulations are carried out to
study the provision of dynamic frequency control
support by energy storage systems [22] in an isolated
system with large shares of RES. A still open issue
consists into integrating evaluations of the impact of
advanced control, protection and defense systems at
a generation planning stage.
In this regard, the present paper aims to
propose a method, based on frequency droop
approach, aimed at fast evaluating the maximum
RES penetration in power systems that are
interconnected to the bulk power system via HVDC
links or a single AC line, as may be the case of
islands, accounting for the degree of deployment of
advanced control, protection and defense systems,
and with special focus on frequency stability which
may be the most critical aspect for system security.
In particular, the investigation aims to show
how the maximum allowable penetration of
renewables depends (1) on the characteristics and
degree of deployment of control, protection, and
defense systems applied to RES, (2) on operational
measures such as HVDC transit, and (3) on specific
components such as batteries; and vice versa, which
measures need to be set up to allow a certain RES
penetration without jeopardizing security.
In the proposed approach, the maximum
RES penetration is evaluated on most constraining,
and rather infrequent situations. As a consequence,
for many hours a year it may be possible to
accommodate the production of a higher RES
capacity. Only for limited periods of time RES
generation will need to be curtailed, leading to
wasted energy and increased system costs.
Economic evaluations of the “optimal” RES capacity
are out of the scope of this work. However, the
proposed algorithm can be easily integrated into
consolidated planning tools which can provide
economic indicators over long time series,
considering a wide range of possible scenarios.
The paper is organized as follows: Section
II presents the methodology for fast assessment.
Section III provides an outline of the model of a real
weakly connected system, i.e. the Sardinian power
system in a 2020 horizon. Section IV compares the
results obtained by cross-validating the static
approach with the time domain simulation of the
most critical contingencies on a dynamic model of
the system. An analysis of the impact of defense and
control systems on penetration of renewables is
reported in Section V. Final conclusions are given in
Section VI.
II. METHODOLOGY FOR
PRELIMINARY ASSESSMENT OF
MAXIMUM RES PENETRATION
RES may impact on all security aspects,
namely frequency, angle and voltage stability, and
on overload. In order to identify critical features
which may limit the penetration of RES in an
isolated or weakly connected system, the following
methodology has been elaborated.
2.1 Critical phenomena and contingencies
RES increase not adequately supported by
network upgrading may create congestions, leading
to overloads and possible cascading tripping with
serious consequences for the system. Voltage control
issues may arise too. However, here it is assumed
that congestion can be solved through “normal”
operational planning and network development.
Similarly, local issues such as voltage support can be
solved by measures that are relatively “easy” to
implement (in terms of time, investment), e.g. the
installation of compensation devices. On the
contrary, when remote areas or islands are
concerned, measures aimed to strengthen the
interconnection to the main grid will be more costly
and lengthy, and may not be viable in practice. This
paper is therefore focused on the issues caused by
weak interconnections.
Engineering experience suggests that the
major criticalities in small and isolated or weakly
connected systems, especially in case of significant
RES generation, are related to the frequency
dynamics resulting from sudden power imbalance.
Accordingly, the most critical contingencies for
frequency stability consist of the loss of power
injections. In particular, as far as isolated systems
are concerned, the under frequency caused by the
loss of a generating unit is typically the most critical
phenomenon. When weakly interconnected systems
are concerned, the most severe contingency for
frequency may be the loss of a connection to the
bulk system, either an HVDC link or an AC line (in
case of single AC interconnection; in case of several
AC interconnection lines, other problems may arise
such as angle stability). In particular, the loss of an
exporting interconnection facility may cause an over
frequency leading to the disconnection of generating
units, in turn causing a power deficit and possibly
blackout. This is more likely in scenarios
characterized by high RES and low load.
2.2 Relevant situations
The maximum criticality for frequency is
determined by the combination of a large

disturbance and a low share of generators under
power/frequency regulation. Accordingly, large RES
injections and low load are critical factors. Two
types of critical situations are identified:
1. nighttime situations with minimum load,
maximum wind power production
2. daytime situations, characterized by high wind
and PV production and load far from peak
The daytime case represents the most
unfavorable situation in terms of load and non-
dispatchable injections, assuming the simultaneous
presence of a sunny hour and a high simultaneity
factor for wind generation, in a spring holiday (i.e.
when neither heating nor cooling are significant on
system loading).
2.3 Security criterion
The security criterion adopted for the
analysis is the N-1, with the following performance
requirement: the post-contingency frequency
deviation must not be greater than 1.5 Hz in the
transient period, and not greater than 0.5 Hz at the
steady state of primary frequency control (i.e. within
30 seconds from the disturbance). This criterion is
consistent with the frequency requirements for
weakly interconnected islands in the European grid
code.
2.4 Regulation droop-based approach
The evaluation of maximum RES
penetration according to frequency performance
requirements can be preliminarily addressed with a
“droop” approach, i.e. according to a simplified
method that does not require a detailed model of the
electrical network (single-busbar representation)
[23]. Broadly speaking, the maximum RES
production is related to the power export through the
interconnections. The approach consists of
identifying the export value through the most critical
interconnector, which is associated to the limit
condition in terms of frequency response in case of
loss of the interconnector, given the available control
resources and other system conditions. To this aim,
transient and permanent droops of primary
frequency regulation are used, as they relate the
frequency deviation, the size of the disturbance, the
capacity of the conventional generators in operation
under primary frequency regulation (Frequency
Containment Reserves, FCR), and the possible
contribution to the regulation coming from DC links,
neighbouring synchronous areas, and RES power
plants (e.g. power modulation controls implemented
in wind turbine and PV generators), according to the
mentioned TSO Grid Code prescriptions [15]. The
approach is briefly described by the following two
steps [24].
In the first step, initial conditions of the
power exported through non-critical HVDC and AC
links are set (these are the links of smaller size with
respect to the largest one). Assumptions on load, sun
and wind allow to determine the base case RES
injections. The parameters of the frequency control
are then considered, namely of conventional
generators, wind generators (if available, at least for
downward regulation), and HVDC systems (for
upward and downward regulation). These quantities,
combined with the requirement of maximum
admissible frequency deviation evaluated for the
reference contingency (e.g. loss of one HVDC pole),
allow to determine the minimum capacity (nominal
power in MW) of conventional units in operation
under primary regulation, which is needed to ensure
the frequency stability of the network, as a function
of the power exported through the critical
interconnector. (It may be recalled that for a given
disturbance, primary regulation contribution is
proportional to the size of the regulating unit.) This
minimum requirement for conventional power in
operation allows to establish a specific configuration
of conventional units to be assumed in service (unit
commitment). These computations are based on the
linearized formulas involving transient and
permanent droop of the primary frequency
regulation. In the resulting unit commitment, the
minimum requirements of secondary and tertiary
reserve must be met as well, otherwise more
generators need to be committed.
Regarding the second step, the selected
configuration of conventional dispatchable units is
characterized by an overall minimum power,
therefore it is necessary to verify the compatibility of
the overall minimum power with respect to the
power balance of the network. In this second step the
amount of RES (wind and PV generation) injected
into the grid can be determined by the following
power balance expression:
orsterconnectinDCorsterconnectinAC
pumpsloadicsphotovoltawindgenalconvention
PP
PPPPP


(1)
where
.... genconvledispatchabnotgenconvledispatchabgenalconvention
PPP 
(2)
With obvious meanings of symbols; the
left-hand side of (1) refers to the generation sources
and the right-hand side to power absorptions (loads
and exports). Thus, if one sets the values for loads,
the greater the injected wind power, the smaller the
injection of conventional power.
In particular, the maximum wind power
injection is obtained by setting dispatchable
conventional plants to their minimum power

production. The greater the number of conventional
plants in service, the greater the resulting overall
minimum allowable production. Under this main
assumption, in order to maximize the wind
penetration one should operate as few as possible
conventional units. On the other hand, the minimum
power production for the conventional thermal units
in service is given by security requirements. In case
of the thermal units described in the test system the
average minimum power is assumed as 40% of
maximum net power output.
Moreover, in order to satisfy the minimum
operating reserve requirements, the selected
dispatchable units can be dispatched to minimum
power output up to 60%. Therefore, in case of a high
lower bound for the minimum allowable production,
the maximum penetration of RES power can be
significantly limited. The minimum operating
reserve includes the Frequency Containment
Reserve (FCR), Frequency Restoration Reserve
(FRR), Replacement Reserve (RR), based on
ENTSO-E Grid Code identification [25] and
corresponding to the TSO Grid Code requirements
as primary reserve, secondary reserve and tertiary
reserve respectively [25][26].
III. TEST SYSTEM
The methodology is applied to an example
of weakly connected power system, namely the
Sardinian system at 2020 horizon. This power
system has a peak load level of about 1000 MW
during low load months and 2000 MW during high
load months. The conventional generation capacity
is composed by 3 GW of fossil-fuelled units and
460 MW of hydro units.
The installed wind capacity in the system is
about 950 MW, of which about a large part
(accounting for 400 installed MW) perform
downward frequency regulation for frequency higher
than 50.3 Hz. The capacity of biomass power plants
is about 77 MW. The 400/220 kV grid is connected
to the continent by two HVDC links: “SACOI” and
“SAPEI” HVDC link. SACOI is a three-terminal
HVDC link with Current Source Converter
technology, monopole 200 kV DC, 300 MW, de-
rated to 150 MW. The minimum power transmission
is 30 MW. SAPEI is a bi-terminal HVDC link with
Current Source Converter technology, bipolar ±500
kV DC, 2x500 MW. The minimum power level is 50
MW for each pole. Sardinia is also AC connected to
Corsica via the “SARCO” 150 kV, 150 MVA
submarine cable.
3.1 Scenario for 2020 horizon
A scenario analysis was performed by
considering hourly load and generation profile for
2020 horizon: the main assumptions are reported in
Table .
Table I. Data for 2020 scenario for sardinian power
system
The hourly load profile was allocated on
local production capacity by profiling the hourly
production on dispatchable and not dispatchable
plants, with priority given to RES according to the
current regulation, and taking into account the transit
constraints on interconnections.
The analysis allowed to identify realistic
situations as “worst cases” from the security
viewpoint, according to the criteria described in
Section 2.2. Two PV penetration scenarios (630 or
800 MW) are considered in 2020 horizon.
3.2 Reserve requirements
The TSO Grid Code requires that in
Sardinia the required FCR must be supplied by all
conventional units with a minimum contribution
equal to ±10% of maximum net power output if the
power unit rating is at least 10 MVA.
Moreover the permanent droop is set to the
default value of 5%. The contribution of wind and
PV units is limited to the power-frequency
modulation in case of over frequency transients
based on Annex 17 and 70 of TSO Grid Code. The
FRR requirements is about ±2÷3% of hourly load
(based on load variation control method proposed by
ENTSO-E). The RR reserve is estimated in line with
the guidelines provided by ENTSO-E [27]: it is
calculated considering the average deviation of the
generation and of hourly load forecast from actual
value ( ADEV ), the standard deviations L
 related
to load forecast error and G
 related to
conventional unit output (to take into account
possible unit outages) as described by (3)
considering a 95% probability ( 645.1 ).
LGADEVRR
22
  (3)
The maximum deviation takes into account
the maximum generation unit loss. In case of
Sardinian system the maximum net power for units
is 200÷300 MW. In presence of wind and PV plants
the above RR expression can be extended including
the standard deviation for the forecast error
associated to the wind and PV production under

main hypothesis of standard Gaussian distribution.
In particular in presence of very high RES
production levels, about 5% and 2% of standard
deviation (normalized on the relevant installed
capacity) are assumed respectively for wind and
solar productions, which is consistent with situations
of high RES production. It is assumed that the most
critical situation that must be compensated by the
RR is the increase of hourly load or decrease of
generation. In fact in case of load decrease or
generation increase it is possible to use the power
modulation control of RES.
3.3 Dynamic models
The dynamic models are developed in
DigSILENT Power Factory to cross-validate the
proposed planning tool. In particular, the models
concern: Control systems (prime movers and
governors of conventional units, primary frequency
control applied to SAPEI HVDC link, over
frequency control on wind farms, over frequency
control on PV plants rated  50 kW, inertia
emulators for wind turbines); Protection systems
(under/over frequency protections for distributed PV
units, over/under speed, over/under voltage
protection for WTs), Defense systems (automatic
load shedding schemes sensitive to frequency and its
derivative, emergency tripping scheme for wind
turbines in case of over frequencies); Loads
(simulating the typical dynamic response of HV
subtransmission networks in Italy); BESS (Battery
Energy Storage system) based on Lithium-ion
technology for grid frequency support.
During cross-validation, the following well-
known indicators are used:
- Post-disturbance steady-state frequency (fss)
- Largest frequency excursion (fextreme)
- Rate Of Change Of Frequency (ROCOF)
IV. SIMULATION RESULTS
The above methodology was applied in two
selected most critical scenarios: a) daily load with
SACOI HVDC link in service and b) daily load
without SACOI link.
Similarly the cross-validation with dynamic
models regard the scenarios reported in Table I.
Table I. RES penetration scenarios for dynamic
simulations
4.1 Scenario 1 – Daily load with SACOI in
service
In this scenario the daytime load of 1200
MW is considered and the HVDC links are both in
service. SACOI is exporting its maximum power.
The analysis of wind penetration is proposed, taking
the export on SAPEI link as a parameter, i.e. the loss
of one pole of the SAPEI HVDC link is the critical
contingency.
The minimum required RR and FRR
reserve is evaluated as 200 MW; the corresponding
conventional units in service is about 680 MW
working at 55% of maximum net power output.
Considering also the must-run thermal units from
biomass and solar thermodynamic technologies
forced at 90% of their maximum net power and 630
MW of PV (without frequency disconnection
thresholds), the droop-based methodology allows to
determine the maximum wind power at minimum
conventional units in service. Fig. 1 shows the
maximum wind penetration assuming 60% of WTs
subjected to over frequency control (without
additional remote defense schemes) and an average
simultaneity factor of 70% (i.e. wind production is
70% of the cumulated nominal installed wind power
in the island). It is possible to install up to 1750 MW
nominal installed wind power taking in service 1000
MW of conventional units (680 MW of dispatchable
units and 335 MW of must-run generators) and 50
MW SAPEI margin. This maximum penetration
level can be achieved with a margin on SAPEI equal
to 50 MW per pole (assuming symmetric operation
of the two poles), i.e. with SAPEI exporting 2 x 450
MW. This result is to be interpreted in this way: up
to a certain extent, an increasing margin on SAPEI
(with respect to the export at maximum level) allows
a higher RES production, as it implies a smaller
perturbation (loss of SAPEI pole) and a post-
contingency higher contribution of primary
frequency regulation from the safe pole of SAPEI (in
the hypothesis of pre-fault symmetric operation).
However, this trend stops at a certain point due to
the presence of non-dispatchable conventional units
and to the need to keep generators in service to
provide reserve (FRR, RR).

The dispatchable units used to guarantee
the operating reserve can be selected among thermal
units (available rated powers 320/160/88 MW). In
this case one possible combination is 2320 MW
fossil-fuelled units.
Fig. 1. Maximum wind power penetration and
minimum nominal power of conventional units
required in operation, daily load + two HVDC links
in service
Fig. 2 shows the time evolution of
frequency in case of the contingency in scenario
S1_1750 for the “worst case” and for different PV
retrofitting levels. Retrofitted PV plants remain
connected to the grid in the whole frequency range
47.5-51.5 Hz, while non-retrofitted ones disconnect
as soon as the frequency exceeds the range 49.7-50.3
Hz.
20,016,012,08,004,000,00 [s]
51,00
50,75
50,50
50,25
50,00
49,75
Bus400kV: Frequency in Sardinia [Hz] 60% WF over freq regulation, no PV retrofit
Bus400kV: Frequency in Sardinia [Hz] 60% WF over freq regulation, 50%PV retrofit
Bus400kV: Frequency in Sardinia [Hz] 60% WF over freq regulation, 60%PV retrofit
Bus400kV: Frequency in Sardinia [Hz] 60% WF over freq regulation, 70% PV retrofit
Bus400kV: Frequency in Sardinia [Hz] 60% WF over freq regulation, 85% PV retrrofit
frequencygrid Date: 5/10/2013
Annex: /12
DIgSILENT
all
Full PV retrofit
Fig. 2. Grid frequency in case of CTG 1 (loss of one
SAPEI pole) for the “worst case” (100% PV
retrofitting and load modelled as PQ constant load)
and for more realistic scenarios of PV retrofitting
(50, 60, 70, and 85%)
The solid curve in Fig. 2 confirms the
outcome of the methodology aimed at assessing the
maximum (“worst case”) wind power penetration. In
fact the contingency determines a steady state
frequency deviation (480 mHz) very close to the
security limit in the Grid Code (500 mHz). The
slight differences are due to discretization issues
present in the more detailed dynamic simulation
(e.g. the choice of specific conventional units
implies a total conventional power in service which
is close to but does not coincide with the one coming
from the static methodology).
In order to prove the correctness of the
proposed approach in evaluating the maximum
installed wind power, the cross-validation process
analyses the S1_1900 scenario which considers an
increment in installed wind power in Sardinia (1900
MW). The same production coefficient is adopted
for the wind farms (70%). The additional wind
power production is compensated by an increase in
SAPEI export (passing from 450 MW to 500 MW
per pole). Also in this scenario 60% of WTs (around
1200 MW of installed wind power) are endowed
with over frequency control.
The dynamic simulation (as well as the
static methodology) assumes a steady state
overloading capacity of 110% for SAPEI link, thus
an additional 50 MW upward reserve is assumed on
each pole. The initial conventional unit production is
unchanged. Fig. 3 shows the frequency in case of
100% PV retrofitting (worst case for over
frequencies) and for more realistic cases with
different PV retrofitting levels. The steady state
frequency deviation (520 mHz) in the worst case
violates the security criteria, in line with the
outcomes of the static methodology.
20,016,012,08,004,000,00 [s]
51,25
51,00
50,75
50,50
50,25
50,00
49,75
Bus400kV: Frequency in Sardinia [Hz] 60% WF over freq reg., all PV retr.
Bus400kV: Frequency in Sardinia [Hz] 60% WF over freq reg., 50%PV retr.
Bus400kV: Frequency in Sardinia [Hz] 60% WF over freq reg, 60%PV retr.
Bus400kV: Frequency in Sardinia [Hz] 60% WF over freq reg, 70% PV retr.
Bus400kV: Frequency in Sardinia [Hz] 60% WF over freq reg, 85% PV retr.
frequencygrid Date: 5/10/2013
Annex: /8
DIgSILENT
Fig. 3. Grid frequency for CTG 1 for the worst case
(all PV retrofitting) and for realistic scenarios of PV
retrofitting
A significant contribution to frequency
regulation comes from wind turbines which undergo
a power reduction ranging from 65 MW (in case of
50% PV retrofitting) to 158 MW (85% PV

retrofitting). The largest frequency deviation ranges
from 50.87 Hz (85% PV retrofitting) to 50.60 Hz
(50% PV retrofitting). The steady state frequency
deviation for 85% PV retrofitting (440 mHz) is still
close to the security limit and further disturbances
can cause security violations. Thus, in the present
case a wind turbine tripping scheme must be adopted
to accommodate the additional wind power installed.
Fig. 4 compares the frequency transient in case of
85% PV retrofitting with (solid line) and without the
emergency tripping scheme with two different
values for tripping threshold, 50.3 Hz and 50.6 Hz
respectively for the dashed and dot-dash line.
20,016,012,08,004,000,00 [s]
51,00
50,80
50,60
50,40
50,20
50,00
49,80
Bus400kV: Frequency in Sardinia [Hz] yes WFtripping 50.3Hz, 85%PV retr.
Bus400kV: Frequency in Sardinia [Hz] no Wf tripping, 85% PV retrofitted
Bus400kV: Frequency in Sardinia [Hz] yes Wf tripping 50.6Hz, 85% PV retr.
frequency3 Date: 5/10/2013
Annex: /11
DIgSILENT
Fig. 4. Grid frequency for 85% PV retrofitting with
(solid line) and without (dashed and dot dash lines)
tripping scheme for WTs
At 50.3 Hz the emergency scheme trips
20% of WTs, thus curtailing around 260 MW
generation. The largest frequency deviation is 50.44
Hz for the 50.3 Hz threshold and 50.67 Hz for the
50.6 Hz threshold, and the steady state frequency
deviation is equal to 180 mHz for both settings
(below the threshold to activate the over frequency
control of WTs).
4.2 Scenario 2 – Daily load without the SACOI
link
The Grid Development Plan of the Italian
TSO takes into account the new SACOI project in
order to replace the existing very old SACOI. This
scenario assumes that at horizon 2020 the existing
SACOI will be almost out of service under the main
hypothesis that the new link will be under
construction. Thus, the scenario excludes the 150
MW export contribution of SACOI link; the
exporting capacity via DC links is reduced to 2500
MW due to SAPEI link. Given the same load level,
reserve requirements and must-run injections, 630
MW of PV installed power imply the possibility to
install up to 1474 MW of wind power (see Fig. 5),
considering a 70 MW margin on each SAPEI pole.
The dynamic simulations are performed on
the specific operating point S0_1475 corresponding
to the peak of the maximum allowable wind power
penetration.
Fig. 5. Maximum wind power penetration, minimum
required conventional power in operation, in case of
daily load and only one HVDC link in service
The steady state frequency after the
application of CTG1 is 50.460 Hz which is close to
the limit imposed by the Grid Code (50.500 Hz): this
confirms the effectiveness of the methodology in
identifying the limiting conditions in terms of
maximum allowable wind power penetration.
V. IMPACT OF CONTROL SYSTEMS
ON MAXIMUM RENEWABLE
PENETRATION LEVEL
Table II reports the results on wind power
penetration cases for daytime and nighttime load
scenarios in presence of over frequency control on
WTs, assuming a total PV installed power equal to
630 MW. Simulations show that a 20% increment in
the share of WTs subject to over frequency control
causes the wind power penetration to pass from 1650
(see Fig. 1, in particular the y-value of the wind
power penetration curve for a SAPEI margin of 20
MW2) to 1850 MW, given the same margin on
SAPEI HVDC link (20 MW2) and the SACOI link
in service. Similar benefits can be found in the
nighttime load scenarios.

Table II. Wind power penetration cases
with daytime and nighttime scenarios, in presence of
over frequency control on a percentage of WTs, and
different status (ON/OFF) of SACOI, 630 MW
installed PV
In case of a higher PV installed power (800
MW in S1_1750_PV800 scenario) the additional
production from PV can be compensated using
available pumped storage which helps accommodate
a higher amount of wind power as Table III shows.
Table III. Wind power penetration cases
with daytime scenarios, in presence of pumping
storage and over frequency control on a share of
WTs, different status (ON/OFF) of SACOI - 800
MW installed PV
To cross-validate these results, the dynamic
simulation focuses on S1_1750_PV800 scenario.
This scenario assumes a daytime load with
SACOI exporting 150 MW and 1750 MW installed
wind power, and a larger PV penetration is
considered (800 MW instead of 630 MW of the base
case S1_1750). Two possible ways to compensate
the additional power output from PVs are analysed:
(1) the additional PV production is exported by
SAPEI poles which have to operate at 500 MW. The
10% overloading capacity is exploited for frequency
regulation; (2) the additional PV production is stored
in a pumping plant (Taloro) with maximum pumping
capability of 240 MW.
Table IV summarises the post-disturbance
steady-state frequency, the largest frequency
excursion and the initial ROCOF found for CTG1
(loss of one SAPEI pole) for the two alternatives,
with reference to the over frequency “worst case”.
Table IV. largest frequency excursion,
initial ROCOF, post-disturbance steady-state
frequency for CTG 1 in the worst case without / with
pumps in service
The first method does not allow to fulfill
security constraint (fss is higher than 50.5 Hz). This
confirms the outcome of the static methodology
which estimates about up to 1560 MW as maximum
wind power to be installed (with a 50 MW margin
on SAPEI poles) in case of 800 MW of installed PV
plants. The use of storage allows to increase the
margin on the SAPEI pole (thus, reducing the power
mismatch if one pole is lost) and it grants a higher
regulating margin on the safe pole.
VI. CONCLUSION
This paper has proposed a fast approach to
evaluate the maximum RES installed power on a
weakly connected power system, and to assess the
benefits of advanced control, defense and protection
systems to increase RES penetration. The relevant
algorithm can be efficiently integrated in generation
planning tools aimed to evaluate the opportunity
(over a long term) of increasing the maximum
renewable penetration in weakly connected power
systems.
The algorithm has been cross-validated on a
2020 model of the Sardinian power system by using
a time domain simulator which runs the most critical
contingency over several scenarios: the matching of
results from the two approaches is good.
The adoption of an “ad hoc” wind tripping
scheme which trips WTs in case of over frequency
transients can allow larger amounts of wind power to
be installed in the grid.
Further work may consist in exploiting the
proposed algorithm to quantify the effect of the
advanced control/defense systems on long term
reliability indicators.
ACKNOWLEDGEMENTS
This work has been financed by the
Research Fund for the Italian Electrical System
under the Contract Agreement between RSE S.p.A.
and the Ministry of Economic Development -
General Directorate for Nuclear Energy, Renewable
Energy and Energy Efficiency - in compliance with
the Decree of March 8, 2006.

REFERENCES
[1]. EWEA (European Wind Energy Association),
Aiming High: Rewarding Ambition in Wind
Energy, November 2015.
[2]. IEA, World Energy Outlook 2015, 2015
[3]. ENTSO-E (European Network Transmission
System Operators Association), Electricity in
Europe 2015, 2016.
[4]. H. Holttinen et al., Design and operation of
power systems with large amounts of wind
power, IEA Wind Task 25, Final report, Phase
three 2012-14, 2016.
[5]. ACER (Agency for the Cooperation of
Energy Regulators), Capacity remuneration
mechanisms and the internal market for
electricity – Article 11 of Regulation (EC) No
713/2009, July 30, 2013.
[6]. ACER/CEER (Council of European Energy
Regulators), Annual Report on the Results of
Monitoring the Internal Electricity and Natural
Gas Markets in 2012, Report, Nov 2013.
[7]. Tradewind, Integrating Wind - Developing
Europe's power market for the large-scale
integration of wind power, Final report,
February 2009.
[8]. European Wind Integration Study (EWIS),
Towards a successful integration of Large Scale
Wind Power into European Electricity Grids,
Final Report, 31 March 2010.
[9]. V. Vittal, J. McCalley, V. Ajjarapu, U. V.
Shanbhag, Impact of Increased DFIG Wind
Penetration on Power Systems and Markets,
PSERC Publication 09-10, October 2009.
[10]. IEA, Harnessing Variable Renewables: a Guide
to the Balancing Challenge, Guide, 2011.
[11]. M. Dernbach, D. Bagusche, S. Schrader,
Frequency Control in Québec with DFIG wind
turbines, 9th International Workshop on Large-
Scale Integration of Wind Power into Power
Systems as well as on Transmission Networks
for Offshore Wind Power Plants Conference,
Québec, Canada, October 18-19, 2010.
[12]. Hydro-Québec, Transmission Provider
Technical Requirements for the Connection of
Power Plants to the Hydro-Québec
Transmission System, February 2009.
[13]. G. Kaestle, T. K. Vrana, Improved requirements
for the connection to the low voltage grid, Proc.
of CIRED, 21st International Conference on
Electricity Distribution, Frankfurt, June 6-9,
2011.
[14]. ENTSO-E, Requirements for Grid Connection
Applicable to all Generators, Grid Code, June
26, 2012.
[15]. ENTSO-E, Automatic frequency disconnection
settings of installed photovoltaic (PV) panels in
some European countries, 18 July 2011.
[16]. ENTSO-E, Assessment of the system security
with respect to disconnection rules of
photovoltaic panels, SPD Report, April 25,
2012.
[17]. ENTSO-E, Dispersed generation impact on CE
region security Dynamic Study, Final Report,
2013
[18]. J. M. O'Sullivan, M. Flynn, M. J. O'Malley,
Modelling of frequency control in an island
power system, Proceedings of the IEEE Power
Engineering Society Winter Meeting, pp. 574 -
579, February 1999.
[19]. L. Sigrist, I. Egido, L. Rouco, Frequency
stability boundary of small isolated Power
Systems, Proc. of IEEE 26-th Convention of
Electrical and Electronic Engineers, Israel,
2010.
[20]. G. Delille, B. François, and G. Malarange,
Dynamic frequency control support: A virtual
inertia provided by distributed energy storage to
isolated power systems, IEEE Innovative Smart
Grid Technologies Europe Conf. (ISGT),
Gothenburg, Sweden, Oct. 2010.
[21]. G. Delille, B. François, and G. Malarange,
Dynamic frequency control support by energy
storage to reduce the impact of wind and solar
generation on isolated power system's inertia,
IEEE Transactions on Sustainable Energy, vol.
3, no. 4, Oct. 2012.
[22]. D. Kottick, M. Blau, D. Edelstein, Battery
energy storage for frequency regulation in an
island power system, IEEE Trans on Energy
Conversion, Vol. 8, Issue 3, Sept 1993, 455-459.
[23]. R. Marconato, Electric power systems, CEI,
Milan, 2002-2008.
[24]. D. Cirio, M. Benini, V. Brignoli, A. Gatti, A.
Gelmini, A. Pitto, Wind power expansion and
security assessment: a 2020 scenario in
Sardinia, MEDPOWER 2012, Cagliari, Italy,
October 1-3, 2012.
[25]. ENTSO-E, Draft Network Code on Load-
Frequency Control and Reserves, Draft version,
May 17, 2016, available at www.entsoe.eu.
[26]. TERNA, Code for transmission, dispatching,
development and security of the grid, last
version 2013, available at www.terna.it.
[27]. ENTSO-E, Supporting Document for the
Network Code on Load-Frequency Control and
Reserves, June 28, 2013.

Assessment of the Dependence of Renewable Penetration Level in a Weakly Connected System on Available Control and Defence Systems

More Related Content

What's hot

Viewers also liked

Similar to Assessment of the Dependence of Renewable Penetration Level in a Weakly Connected System on Available Control and Defence Systems

Recently uploaded

Assessment of the Dependence of Renewable Penetration Level in a Weakly Connected System on Available Control and Defence Systems